SYSTEMATICS OF NECROMYS (RODENTIA, CRICETIDAE,SIGMODONTINAE): SPECIES LIMITS AND GROUPS, WITHCOMMENTS ON HISTORICAL BIOGEOGRAPHY

GUILLERMO D’ELIA,* ULYSES F. J. PARDINAS, J. PABLO JAYAT, AND JORGE SALAZAR-BRAVO

Departamento de Zoologıa, Facultad de Ciencias Naturales y Oceanograficas, Universidad de Concepcion,Casilla 160-C, Concepcion, Chile (GD)Centro Nacional Patagonico, CC 128, 9120, Puerto Madryn, Chubut, Argentina (UFJP)Laboratorio de Investigaciones Ecologicas de las Yungas (LIEY), Facultad de Ciencias Naturales eInstituto Miguel Lillo, UNT, CC 34, 4107 Yerba Buena, Tucuman, Argentina (JPJ)Department of Biological Sciences, Texas Tech University, Lubbock, TX 79409, USA (JSB)

We present the most comprehensive systematic study to date of Necromys, a rodent genus distributed in open

areas north and south of Amazonia and in Andean grasslands. The study is based on sequences of the

cytochrome-b gene that were analyzed by parsimony and Bayesian approaches. The analyses include sequences

of 62 specimens from 51 localities from Argentina, Bolivia, Brazil, Colombia, Paraguay, Peru, Uruguay, and

Venezuela, representing all but 1 of the species currently recognized in the genus. Necromys was recovered as

a monophyletic group and we found a large polytomy at its base that involves 3 lineages. One, represented by the

Andean N. lactens, shows a marked phylogeographic pattern. The 2nd clade is formed by N. urichi from the

northern grasslands of South America and N. amoenus from the central Andes. Results suggest that each of these

taxa may represent more than 1 biological species. The 3rd clade is formed by lowland species found south of

Amazonia. Within this clade N. obscurus is sister to the remaining species. Haplotypes recovered from specimens

assigned to N. benefactus, N. temchuki, and N. lasiurus form a clade, but these taxa do not form reciprocally

monophyletic groups, nor does this large clade possess geographic structure. These genealogical results,

discussed in the context of genetic variation, are the basis of taxonomic (e.g., N. benefactus and N. temchuki are

regarded as junior synonyms of N. lasiurus) and biogeographic considerations.

Key words: Bolomys, Cabreramys, Cricetidae, gene tree, Muroidea, Necromys, South America, systematics, taxonomy,

Zygodontomys

The diversity of living sigmodontines is impressive; at last

count recent species numbered 380 in about 84 genera (D’Elıa

et al. 2007; Musser and Carleton 2005; see also D’Elıa and

Pardinas 2007). These numbers make Sigmodontinae the most

diverse subfamily of New World mammals and on a global

scale 2nd only to the Old World Murinae. Diversity of species

is unevenly distributed among sigmodontine genera, with

Akodon and Thomasomys having well over 40 species, whereas

many genera (e.g., Deltamys, Lundomys, and Sooretamys) are

monotypic, presenting the well-known (although not well-

understood) hollow-curve distribution of species richness (Reig

1986). Among the most diverse genera in the akodontine tribe,

Necromys includes 9 recognized species (Musser and Carleton

2005).

Necromys departs from other sigmodontine genera of

equivalent diversity (e.g., Abrothrix, Neacomys, and Bruce-pattersonius) because of its distribution. It occurs in open areas

from Venezuela and Trinidad and Tobago in the northern part

of South America to central Argentina in the south, and from

the Atlantic in Brazil, Uruguay, and Argentina to Andean

elevations of about 5,000 m above sea level. Moreover,

Necromys is the 2nd oldest known South American sigmo-

dontine genus (2nd only to Auliscomys), because N.bonapartei� is known from sediments of about 4–3.5 million

years of age, corresponding to the Chapadmalalan of the late

Pliocene (Pardinas and Tonni 1998). Therefore, strong

hypotheses about relationships among species in the genus

and the position of Necromys within Akodontini are of

fundamental importance to understanding sigmodontine sys-

tematics and biogeography.

* Correspondent: guillermo@udec.cl

� 2008 American Society of Mammalogistswww.mammalogy.org

Journal of Mammalogy, 89(3):778–790, 2008

778

The phylogenetic position of Necromys within the sigmo-

dontine radiation is well established; it has been addressed by

the analysis of both mitochondrial and nuclear gene sequences

(D’Elıa 2003; Smith and Patton 1999). Necromys forms part of

the Akodon division of Akodontini, and is sister to Thalpomys(D’Elıa 2003).

Necromys is different from most other genera of sigmodon-

tines because it was originally described from a fossil form

(Ameghino 1889). Species currently recognized in Necromyshave been placed in Akodon, Bolomys, Calomys, Cabreramys,

Chalcomys, or Zygodontomys (Table 1; for a complete review

of the taxonomic history of Necromys see Anderson and Olds

[1989], Reig [1987], and Tate [1932]). The current contents of

Necromys were established 30 years ago, when Reig (1978)

resurrected Bolomys to encompass a group of species

morphologically similar to amoenus, type species of the genus

(Thomas 1916). Some of the species that Reig (1978, 1987)

placed in Bolomys had been, until then, considered part of

Akodon and Zygodontomys (e.g., amoenus, lactens, and

lasiurus), whereas others (e.g., benefactus, obscurus, and

temchuki) were in Cabreramys erected by Massoia and Fornes

(1967). Like Reig (1978), Massoia and Fornes (1967)

TABLE 1.—Abridged historic account of living taxonomic forms assigned to Necromys.

Gyldenstolpe (1932) Cabrera (1961) Reig (1987)a Musser and Carleton (2005) Present study

Bolomys amoenus Akodon (B. amoenus) B. amoenus Necromys amoenus N. amoenus (Thomas, 1900)

B. lactens A. (B.) lactens (includes

negrito orbus)

B. lactens (includes

negrito orbus)

N. lactens (includes

negrito orbus)

N. lactens (Thomas, 1918)

(includes negrito (Thomas,

1926), orbus (Thomas, 1919))

B. negrito

B. orbus

Zygodontomys lasiurus Z. lasiurus (includes fuscinus,

brachyurus, pixuna)

B. lasiurus (includes

arviculoides, brachyurus,fuscinus, lasiotis, pixuna)

N. lasiurus (includes

arviculoides, brachyurus,fuscinus, lasiotus,

orobinus, pixuna,

renggeri)

N. lasiurus (Lund, 1840)

(includes arviculoides(Wagner, 1842), benefactus

(Thomas, 1919), brachyurus

(Wagner, 1845), elioi

(Contreras, 1982), fuscinus(Thomas, 1897), lasiotus

(Lund, 1839), liciae

(Contreras, 1982),

orobinus (Wagner, 1842),

pixuna (Moojen, 1943),

renggeri (Pictet, 1844),

temchuki (Massoia, 1982))

A. arviculoides (includes

orobinus)

A. arviculoides (includes

orobinus, renggeri)

A. benefactus Included in A. obscurus Included in B. obscurus N. benefactus

B. temchuki (includes

elioi, liciae)

N. temchuki (includes

elioi, liciae)

Hesperomys brachyurus

Z. fuscinus

A. lenguarum Included in A. obscurus B. lenguarum (includes

tapirapoanus)N. lenguarum (includes

tapirapoanus)N. lenguarum (Thomas, 1898)

(includes tapirapoanus

(Allen, 1916))

Z. tapirapoanus A. tapirapoanusA. obscurus A. obscurus B. obscurus (includes

benefactus)

N. obscurus (includes

scagliarum)

N. obscurus (Watherhouse,

1837) (includes scagliarum

Galliari and Pardinas, 2000)B. innom. sp. (would

correspond to the later

scagliarum)

Z. punctulatus Z. punctulatus Not considered N. punctulatus N. punctulatus (Thomas, 1894)

A. (Chalcomys) urichi A. urichi (includes

chapmani, meridensis,

saturatus, venezuelensis)

A. urichi (includes

saturatus, venezuelensis)

N. urichi (includes

chapmani, meridensis,

saturatus, tobagensis,

venezuelensis)

N. urichi (Allen and Chapman,

1897) (includes chapmani(Allen, 1913), meridensis

(Allen, 1904), saturatus

(Tate, 1939), tobagensis(Goodwin, 1962),

venezuelensis (Allen, 1899))

A. (Ch.) chapmani

A. (Ch.) meridensisA. (Ch.) venezuelensis

a Some taxa are missing from Reig’s (1987) scheme because he only dealt with taxa he considered to be Akodontini; moreover, because his work was not a taxonomic catalogue he did

not mention all akodonts.

June 2008 779D’ELIA ET AL.—SYSTEMATICS OF NECROMYS (SIGMODONTINAE)

recognized that some species placed in Akodon and Zygodont-omys constituted a homogenous group distinct from the typical

morphotypes of both genera. However, Massoia and Fornes

(1967) failed to acknowledge that the species in consideration

also were associated with amoenus, type species of Bolomys,

which was available to contain those forms. Therefore, Reig

(1987) placed Cabreramys under the synonymy of Bolomys.

Massoia and Pardinas (1993) recognized that Necromysconifer�, a fossil form described by Ameghino (1889) from

the late Pleistocene of the Argentinean Buenos Aires Province

and type species of Necromys, is a synomym of B. benefactus.

Because Necromys has priority over Bolomys (and Cabrer-amys), this is the generic name currently used for the group of

akodontine mice herein studied. Further, Smith and Patton

(1993) expanded Necromys by showing that Akodon urichi,from northern South America, belonged to Necromys. This

hypothesis was later corroborated with broader taxonomic

(Smith and Patton 1999) and character (D’Elıa 2003) sampling.

Galliari and Pardinas (2000) suggested that Bolomys may be

retained to encompass the Andean species and Necromysrestricted to lowland forms. D’Elıa (2003) discussed this

hypothesis together with other classificatory alternatives,

favoring the broad definition of Necromys that is reflected in

current classifications (e.g., Musser and Carleton 2005). Today,

9 species are placed in Necromys (Musser and Carleton 2005).

Nevertheless, and despite the abundance of specimens in

museums of natural history, the alpha taxonomy of the genus

remains chaotic and species boundaries remain poorly un-

derstood. In part, this is because most revisionary works have

been conducted at local scales or within a single species

(Anderson and Olds 1989; Galliari and Pardinas 2000; Macedo

and Mares 1987; Oliveira 1992).

The goal of this study is to generate a phylogenetic

hypothesis for Necromys. This genealogy allows us to evaluate

the content of the genus and its species limits, and to pose

preliminary comments on its historical biogeography.

MATERIALS AND METHODS

Taxonomic sampling and sequence acquisition.—Animal

care and use procedures followed guidelines approved by the

American Society of Mammalogists (Gannon et al. 2007).

Genealogical analysis was based on the first 801 base pairs (bp)

of the cytochrome-b gene of 62 specimens of Necromysbelonging to 50 populations from Argentina, Bolivia, Brazil,

Colombia, Paraguay, Peru, Uruguay, and Venezuela (Fig. 1;

Appendix I). Specimens from which DNA sequences were

gathered belong to 8 of the 9 currently recognized species of

Necromys. We are missing representatives of N. punctulatus,

a species with uncertain geographic distribution in Ecuador

and Colombia. However, our taxonomic coverage is broad; it

includes specimens assignable to several of the nomimal forms

(e.g., Akodon fuscinus, Bolomys temchuki liciae, and Necromysobscurus scagliarum) currently considered synonyms of valid

species, as well as specimens collected at or near some type

localities of several forms (e.g., Akodon chapmani, A. fuscinus,

A. lactens, Cabreramys temchuki, Mus lasiurus, and N. ob-

scurus scagliarum). The phylogeny was rooted with haplotypes

of specimens of Akodon, Deltamys, Thalpomys, and Thap-tomys. Thalpomys is the sister genus of Necromys and all 5

genera constitute the Akodon division (sensu D’Elıa 2003) of

the tribe Akodontini. Sequences of 11 Necromys specimens

and those of the outgroups were obtained from Smith and

Patton (1991, 1999), Anderson and Yates (2000), D’Elıa

(2003), and D’Elıa et al. (2003). Cytochrome-b gene sequences

generated in this study were amplified and sequenced from

a single fragment using primers MVZ 05–MVZ 16 (da Silva

and Patton 1993) and the protocol detailed in D’Elıa and

Pardinas (2004). In all cases, both heavy and light DNA strands

were sequenced and compared. Sequences were visualized,

reconciled, and translated to proteins to proof for stop codons

using Sequence Navigator version 1.0.1 (Applied Biosystems,

Inc., Foster City, California) and SeqMan (DNASTAR 2003).

All sequences were deposited in GenBank (accession numbers

EF531643–EF531693).

Analyses.—Sequences were aligned with Clustal X

(Thompson et al. 1997) by using the default values for all

FIG. 1.—Map of collecting localities of the specimens of Necromysused in the present study. Locality numbers refer to Appendix I. The

type localities of the Necromys species considered valid in Musser and

Carleton (2005) also are shown. The exact type locality of N.punctulatus is not known and is here depicted with a question mark in

Ecuador, the country from which it is supposed to have come.

780 JOURNAL OF MAMMALOGY Vol. 89, No. 3

alignment parameters; no adjustment by eye was needed. Ob-

served percentage of sequence divergence was calculated with

PAUP* (Swofford 2000) ignoring those sites with missing

data. Aligned sequences were subjected to maximum parsimo-

ny (Farris 1982; Kluge and Farris 1969) and Bayesian

inference (reviewed in Huelsenbeck et al. 2001). Characters

used in maximum-parsimony analysis were treated as un-

ordered and equally weighted. PAUP* (Swofford 2000) was

used to perform 200 replicates of heuristic searches with

random addition of sequences and tree-bisection-reconnection

branch swapping. Two measures of clade support were cal-

culated. Because of computational time, Bremer support values

(BS—Bremer 1994) were computed for selected nodes in

PAUP* using command files written in TreeRot version 2

(Sorenson 1999) and with MAXTREE set to 10,000. One

thousand parsimony jackknife (JK—Farris et al. 1996) repli-

cations with 1 addition sequence replicate each, the deletion of

33% of the character data, and MAXTREE set to 1,000 were

performed using PAUP*. Branches with ,50% of support

were allowed to collapse. Bayesian analysis was conducted in

MrBayes 3.1 (Ronquist and Huelsenbeck 2003). Ten analyses

each consisting of 2 independent runs with 3 heated and 1

cold Markov chains each were performed. For each analysis,

a model with 6 categories of base substitution, a gamma-

distributed rate parameter, and a proportion of invariant sites

was specified; all model parameters were estimated in MrBayes.

Uniform interval priors were assumed for all parameters except

base composition and GTR parameters, which assumed

a Dirichlet process prior. Runs were allowed to proceed for 1,

1.5, 2, and 2.5 million generations; trees were sampled every

100 generations for each chain. To check that each run

converged on a stable log-likelihood value, we plotted the

log-likelihood values against generation time for each. The first

25% of the trees were discarded as burn-in and the remaining

trees were used to compute a 50% majority rule consensus tree

and obtain posterior probability (PP) estimates for each clade.

For N. amoenus, a hierarchical analysis of the distribution of

genetic diversity was conducted in the form of analyses of

molecular variance (AMOVA—Excoffier et al. 1992) using

Arlequin version 2.000 (Schneider et al. 2000). Hierarchical

levels were defined on the basis of sampling localities and

major clades found in the phylogenetic analysis. Similarly, for

the sample of N. lasiurus the neutrality tests of Tajima (1989)

and Fu (1997) were implemented in Arlequin.

RESULTS

The data set contained 348 variable sites, 69 of which

correspond to 1st codon positions, 45 to 2nd positions, and 234

to 3rd positions. Observed sequence divergence within and

between localities of the same species ranges from 0% to

3.87% (based on specimens 9 localities of N. amoenus, N.lasiurus, N. lenguarum, and N. obscurus) and 0% to 10.86%,

respectively. Values of comparisons between species pairs

range from 5.2% to 18.26%.

Of the 348 variable characters of the whole data set, 263 are

parsimony informative. The parsimony analysis of this data set

produced 22,107 shortest trees, each 982 steps long. Each tree

had an ensemble consistency index (CI) of 0.473 and a retention

index (RI) of 0.798. The strict consensus tree (Nelson 1979)

defined 44 nodes belonging to Necromys (Fig. 2), 10 of which

involved basal polytomies. At the base of the Necromys clade

are 3 lineages: a clade composed of N. urichi and N. amoenus(JK ¼ 54, BS ¼ 1); another clade formed by N. lactens (JK ¼100, BS ¼ 20); and a clade formed by all lowland species south

of Amazonia (JK ¼ 72, BS ¼ 1). Support for the clades

recovered in the strict consensus is highly variable. In the

jackknife analysis 3 nodes of the Necromys clade, involving

haplotypes of N. lasiurus, collapse at the cutoff of 50%.

Bremer support values for nodes of the Necromys clade range

from 1 to 22.

The Bayesian analyses produced 3 different topologies,

where the relationships among the groups of species varied. For

example, N. lactens alternates from being the sister group to the

remaining Necromys (PP for this later clade ¼ 0.50) in 1- and

2-million generation runs, to being the sister to the ‘‘southern

lowland’’ clade (PP for this relationship , 0.55) in 1-, 1.5-, and

2-million generation runs. In addition, in 2.5-million generation

runs Bayesian analysis recovered the same basal polytomy

found in the maximum-parsimony analysis. In light of these

results, we have chosen to discuss only the Bayesian topology

with the basal polytomy (Fig. 3). This tree shows a total of 12

intraspecific polytomies.

DISCUSSION

Necromys is a sigmodontine genus in great need of revision.

Problems range from nomenclatural issues, to species limits, to

the delimitation of the genus itself (Anderson and Olds 1989;

Galliari and Pardinas 2000; Reig 1987). Few recent studies

have tackled these issues. Galliari and Pardinas (2000) studied

the southern subtropical lowland species based on qualitative

and quantitative morphologic variation. Ventura et al. (2000)

used quantitative morphologic variation to assess the status of

different taxonomic forms of N. urichi. Provensal et al. (2005)

assessed the genetic and morphometric variation of 4 popu-

lations of N. benefactus. Thus, the present study is the 1st one

dealing with systematics of Necromys by analyzing most of the

recognized species with a broad geographic coverage and an

explicit genealogical approach based on variation of the

cytochrome-b gene. Maximum-parsimony analysis retrieved

22,107 shortest trees of 982 steps. Shortest trees have relatively

high consistency (0.473) and retention (0.798) indices. Their

strict consensus is a well-resolved cladogram, although clades

show a wide range of support (Fig. 2). Bayesian analysis (Fig.

3) retrieved a topology congruent for the most part with that of

the maximum-parsimony analysis. A few clades appearing in

the Bayesian consensus are collapsed in the parsimony

consensus; the reverse is also true with regards to a few

clades. Discrepancies between both topologies mostly involve

intraespecific relationships with low support in one or other

consensus tree. The following discussion is structured along

some of the main problems of Necromys systematics and on

clades recovered by our analyses.

June 2008 781D’ELIA ET AL.—SYSTEMATICS OF NECROMYS (SIGMODONTINAE)

The basal radiation of Necromys.—Necromys is recovered

as monophyletic in both analyses, although with deeply

different levels of support. In the maximum-parsimony anal-

ysis, Necromys received 53% of JK support, and only 1 extra

step is needed to collapse the clade and refute the hypothesis of

monophyly; on the contrary, in the Bayesian analysis a mono-

phyletic Necromys is strongly supported (PP ¼ 0.96). This

difference is notable, but not necessarily unexpected. Empirical

studies by Simmons et al. (2004) have shown that JK tends to

underestimate support values, that the Bayesian method con-

sistently overestimates support, and that both methods perform

poorly when the taxon sample is large but the character sample

is relatively limited. It is likely that the latter may explain our

disparate results, because we sequenced more than 60 animals,

but examined only a fragment of the cytochrome-b gene. In

addition, theoretical and simulation studies lack consensus on

the merit of using PP as a support measure in phylogenetics

(reviewed in Alfaro and Holder [2006]). In a maximum-

parsimony analysis of combined mitochondrial (complete

cytochrome-b) and nuclear sequences (759-bp fragment of

the 1st intron of the interphotoreceptor retinoid-binding protein

[IRBP] gene) with a broad sampling of sigmodontines, D’Elıa

FIG. 2.—Strict consensus tree of the 22,107 most-parsimonious trees (length 982, CI ¼ 0.473, RI ¼ 0.798) obtained in the maximum-

parsimony analysis of the cytochrome-b gene sequences. Numbers indicate parsimony jackknife (left of the diagonal; only values .50% are

shown) and Bremer support (right of the diagonal) values of the nodes to their right. If only a number appears it corresponds to the jackknife value.

A node without any number at its left implies that the node has less than 50% of jackknife support and that no Bremer support was calculated for

it. Numbers next to specimen identification refer to the locality where they were collected (see Fig. 1 and Appendix I). An asterisk (*) or a pound

symbol (#) indicates that before this study the specimen in question was assigned to Necromys benefactus or N. temchuki, respectively.

782 JOURNAL OF MAMMALOGY Vol. 89, No. 3

(2003) showed that monophyly of Necromys was strongly

supported (JK ¼ 99, BS ¼ 9). However, that study lacked

several species in the genus, including representatives of N.lactens, 1 of the 3 main lineages recovered in our analyses

above. Several studies (e.g., Galewski et al. 2006; Steppan

et al. 2005) have shown that the utility of mitochondrial

sequences declines rapidly with increasing phylogenetic depth.

Thus, until nuclear sequences for more species of Necromys are

available, it is difficult to reconcile the disparity between

support values in this study and those presented elsewhere.

Analyses reported in Figs. 2 and 3 found a polytomy at the

base of Necromys, composed of the following 3 lineages:

a clade composed of N. urichi and N. amoenus (JK ¼ 54, BS ¼1, PP ¼ 0.93); another formed by N. lactens (JK ¼ 100, BS ¼20, PP¼1.00); and a clade formed by all lowland species south

of Amazonia (JK ¼ 72, BS ¼ 1, PP ¼ 1.00). Considering this

polytomy, an alternative classification option is to recognize 3

genera: Necromys for the morphologically homogeneous

lowland forms; Bolomys for amoenus and urichi; and a 3rd

genus to encompass lactens. We do not agree with this

alternative for 2 reasons: phenotypic patterns of morphological

similarity show that N. lasiurus and N. lenguarum are more

similar to N. lactens than to N. obscurus (Galliari and Pardinas

2000), although in our analyses obscurus is shown to be a sister

FIG. 3.—Majority-rule consensus resulting from the Bayesian analysis of the cytochrome-b gene sequences. Numbers indicate posterior

probability values of the nodes at their right. Numbers next to specimen identification refer to the locality where they were collected (see Fig. 1 and

Appendix I). An asterisk (*) or a pound symbol (#) indicates that before this study the specimen in question was assigned to Necromys benefactusor N. temchuki, respectively.

June 2008 783D’ELIA ET AL.—SYSTEMATICS OF NECROMYS (SIGMODONTINAE)

to a clade formed by lasiurus and lenguarum; and the generic

diagnosis of Necromys (Anderson and Olds 1989) accommo-

dates the discrete levels of variation among the recognized

species, including amoenus. Given the data currently available,

we agree with the contents and limits of Necromys as currently

accepted (i.e., Musser and Carleton 2005). Nevertheless, this

hypothesis should be further tested with the inclusion of N.punctulatus and analysis of nuclear loci.

Necromys lactens as 1 of the 3 main lineages ofNecromys.—Necromys lactens (JK ¼ 100, BS ¼ 20, PP ¼1.00) represents a highly differentiated lineage within the

radiation of Necromys. It ranges from northwestern Argentina

to central Bolivia at elevations between 1,500 and 4,000 m.

This is a region with a complex topography and associated

biotic formations. In northern Argentina and southern Bolivia,

N. lactens occurs on high-altitude grasslands that develop

above the ‘‘sierras’’ (Jayat and Pacheco 2006). The sierras are

a complex of parallel-running hills of different heights and

extensions that replace themselves from east to west and north

to south. The grasslands of the sierras are separated from each

other by relatively large river valleys covered with arid or

semiarid habitats (e.g., Chaco, Monte, and Prepuna) and the

humid Yungas forest (Cabrera 1976).

We included haplotypes recovered from specimens collected

at or near the type localities of Akodon lactens Thomas, 1918

(JPJ 630), A. orbus Thomas, 1919 (JPJ 419), and Bolomysnegrito Thomas, 1926 (JPJ 447). Currently, the latter forms are

considered synonyms of N. lactens. Our results support this

taxonomy: these haplotypes appear closely related and the

percentage of divergence among them is around 2.0%.

Our sampling of N. lactens was not large enough to

constitute a phylogeographic analysis, but our results provide

a preliminary view on the geographic structure of this species.

We sequenced 7 specimens, each from a different locality in

Argentina, including 1 close to the type locality. All have

different haplotypes and genetic differences among them range

from 0.1% to 3.8%. Genetic variation within N. lactens is

geographically structured: a northern clade (JK ¼ 95, BS ¼ 2,

PP ¼ 0.93) is formed by 2 haplotypes from the north of Salta

Province; and a central-southern clade (JK ¼ 91, BS ¼ 2, PP ¼0.93) is represented by samples from Catamarca, Jujuy (south),

and Tucuman. Genetic divergence between clades ranges from

3.0% to 3.8%. Specimens from the northern clade were

collected on the western slopes of the Sierra de Santa Victoria,

near the Bolivian border, and Sierra de Zenta. These systems

are separated from the central and southern sierras by the

Grande River that runs along the Quebrada de Humahuaca.

Haplotypes within the central-southern clade also are geo-

graphically arranged. Those from the farthest south (Sierras de

Ambato, Ancasti, and Cumbres Calchaquies systems) form

a well-supported clade (JK ¼ 98, BS ¼ 3, PP ¼ 1.00) that in

the maximum-parsimony analysis is sister to a central clade

(JK ¼ 71, BS ¼ 2; Chani and Centinela systems).

Our results suggest that the phylogeography of N. lactenswas shaped by the history of high-altitude grasslands in the

central Andes and associated mountain systems. Today,

glaciers occupy some of the mountain ridges in the area (e.g.,

Nevados del Aconquija, Nevados del Chani, and Nevados de

Cachi) above 5,000 m. During the ice ages, glaciers in these

areas extended to 4,000 m, and even to 3,800 m in some places

(Halloy 1978; Rohmeder 1942). This must have lowered the

vegetation belts on mountain slopes and increased the extent of

vegetative zones at higher elevations. Under these conditions,

current patches of high-altitude grassland inhabited by N.lactens may have come into contact and allowed dispersal of

the species along the different sierras. This hypothesis is

supported by the discovery of a fossil community of Lujanian

Age (late Pleistocene) in which high-altitude grassland species

occur at a relatively low altitude (1,900 m). This reinforces the

idea that at least some members of present-day communities

were displaced to lower elevations in the Pleistocene (Ortiz

et al. 1998; see also Ortiz et al. 2000). In light of these results,

a broad phylogeographic analysis of N. lactens and other

grassland specialists (e.g., Phyllotis osilae—Jayat and Pacheco

2006) would be of much interest.

The clade of N. amoenus and N. urichi.—Necromysamoenus and N. urichi form a natural group in both analyses

but the group received strong support only in the Bayesian

analysis (PP ¼ 0.93). In a combined maximum-parsimony

analysis of mitochondrial and nuclear sequences that included

4 species of Necromys, N. amoenus and N. urichi formed

independent lineages in a polytomy at the base of the genus

(D’Elıa 2003). The N. amoenus-N. urichi clade has a large

distribution in the central and northern Andes, extending from

northwestern Argentina to Trinidad and Tobago, and including

the highlands of Bolivia, Colombia, and Peru, and lowlands of

Venezuela.

Based on geography, N. punctulatus—the only currently

recognized species of Necromys not included in our analyses—

may be part of or closely related to the N. amoenus-N. urichiclade, although we note that N. punctulatus is morphologically

indistinguishable from N. lasiurus (Voss 1991). The assumed

distribution of N. punctulatus is Ecuador and probably

Colombia (Musser and Carleton 2005), although several

attempts to obtain additional specimens of this species in the

last 15 years in various habitats in Ecuador have failed.

The last expansion of content in Necromys was proposed by

Smith and Patton (1999), who placed in it the form urichi. This

species ranges throughout the islands of Trinidad and Tobago,

but it also occurs in Colombia and Venezuela. The taxonomy

of N. urichi is presently unsettled. One of the main conflicts

centers on whether the form saturatus from Auyan-Tepui in

southern Venezuela represents a subspecies of N. urichi(Ventura et al. 2000) or a separate species (Linares 1998). At

least 2 other forms, meridensis and venezuelensis, are

considered valid subspecies (Ventura et al. 2000).

In our analyses, we used 4 specimens assigned to N. urichi: 2

from Venezuelan localities south of the Orinoco (1 geo-

graphically proximate to the type locality of saturatus), 1 from

the Caribbean coast of Venezuela (geographically proximate to

the type locality of venezuelensis), and 1 from the Andes of

Colombia (proximate to the type locality of chapmani). The

4 specimens have highly divergent haplotypes (genetic dis-

tances 3.6–9.0%) that formed a well-supported clade (JK ¼ 98,

784 JOURNAL OF MAMMALOGY Vol. 89, No. 3

BS ¼ 6, PP ¼ 1.00). Relationships among the 4 haplotypes are

fully resolved and highly supported in the maximum-

parsimony analysis (Fig. 2), although 1 of the clades collapses

in the Bayesian analysis (Fig. 3). Haplotypes from specimens

collected south of the Orinoco, which can be ascribed to

saturatus (Linares 1998; Ventura et al. 2000), do not form

a monophyletic group. In the maximum-parsimony analysis, 1

of these haplotypes from nearby San Ignacio de Yuruani falls

in a well-supported clade with a haplotype from the Caribbean

coast that represents venezuelensis (JK ¼ 96, BS ¼ 2). Both

localities are less than 1,000 m above sea level. Of the 4

haplotypes studied, these 2 are the most similar, showing 3.6%

divergence. However, this clade collapses in the Bayesian

analysis. The 2nd haplotype assignable to saturatus comes

from southern Venezuela (Cerro de la Neblina, near the

Brazilian border). This is the most divergent of all haplotypes

(genetic distances 8.1–9.0%) and is recovered as sister to the

remaining haplotypes of N. urichi (JK ¼ 96, BS ¼ 4, PP ¼0.99). These results cast doubt on the genealogical distinctness

of saturatus, at least as it is currently delimited (Ventura et al.

2000). The specimen from Cerro de la Neblina was collected at

2,100 m, an elevation similar to that of the type locality

(Auyan-Tepui) of saturatus. Perhaps venezuelensis ranges at

low elevations south and north of the Orinoco, with saturatusoccurring at higher altitudes. If so, venezuelensis, saturatus,

and chapmani (another highland form that appears sister to

venezuelensis) should be given equivalent rank (i.e., species or

subspecies). Regardless of the status of saturatus, our results

imply that the current geographic delimitation of this taxon is

erroneous and illustrate the need to reevaluate the status of

forms currently assigned to N. urichi. As currently defined, this

species may be a composite of .1 biological species.

However, any taxonomic changes at this point are premature.

Five specimens of N. amoenus were analyzed and a different

haplotype was recovered from each; the monophyly of the

species appears well supported in both analyses (JK ¼ 100, BS

¼ 10, PP ¼ 1.00). Interestingly, N. amoenus shows a deep

phylogeographical break between haplotypes recovered from

specimens collected in Cuzco and Puno, in southern Peru (JK

¼ 100, BS ¼ 8, PP ¼ 1.00) and those from Tarija in southern

Bolivia and Salta in northwestern Argentina (JK ¼ 100, BS ¼22, PP ¼ 1.00). Genetic divergence between these clades is

large, ranging from 7.75 to 10.9%. Moreover, an AMOVA

indicates that 88.1% of the genetic variation recovered in the

sample of N. amoenus is due to differences among the 2 main

clades. This suggests that N. amoenus as now understood may

be a composite of .1 biological species. To test this hypothesis

more individuals need to be analyzed, in particular from central

Bolivia, and a comprehensive morphological study undertaken.

A clade of species from the open lowlands south of theAmazon.—A large clade formed by haplotypes from all species

of Necromys from the lowlands south of Amazonia (i.e., N.benefactus, N. lasiurus, N. lenguarum, N. obscurus, and N.temchuki) was recovered with high support in both analyses

(JK ¼ 72, BS ¼ 1, PP ¼ 1.00). However, not all lowland

species were recovered as monophyletic. The 1st split within

the lowland clade separates individuals assigned to N. obscurus

from the remaining species (Figs. 2 and 3). This result

corroborates previous analyses of morphological variation

showing that N. obscurus is markedly different from the other

lowland species (Galliari and Pardinas 2000; Massoia and

Fornes 1967). Support for the monophyly of N. obscurus is

high (JK ¼ 100, BS ¼ 12, PP ¼ 1.00). Two subspecies with

disjunct distributions are currently recognized in N. obscurus.

The nominal form ranges along a narrow band of habitat

bordering the La Plata River and the Atlantic coast of Uruguay,

whereas N. o. scagliarum is known from only a few localities

centered on the Argentinean city of Mar del Plata (Galliari and

Pardinas 2000). In the maximum-parsimony analysis, N. o.scagliarum was recovered as monophyletic (JK ¼ 64), but in

the Bayesian analysis it was paraphyletic with respect to N. o.obscurus. Genetic distances among subspecies average 1.0%,

comparable to the distance among haplotypes of N. o.scagliarum (0.9%). In contrast, both specimens of N. o.obscurus, collected in localities about 80 km apart, share the

same haplotype. More specimens need to be analyzed to test

whether N. o. scagliarum has more genetic variation than N. o.obscurus or whether the observed pattern is caused by our

limited geographic sampling.

Our study also corroborates that at least 2 species of

Necromys inhabit the Chaco, one of the most poorly studied

ecoregions of South America. One of the species is N. lasiurus(JK ¼ 99, BS ¼ 7, PP ¼ 1.00), whose distribution encom-

passes a myriad of ecological regions. For convenience, the

2nd Chacoan species is here referred to as N. lenguarum (JK ¼100, BS ¼ 10, PP ¼ 1.00). This taxon is currently considered

a valid species, but at times it has been regarded as a synonym

of N. lasiurus (e.g., Macedo and Mares 1987). On average,

haplotypes of the 2 species differ by 5.4% (range 5.2–8.1%)

and the monophyly of each species is strongly supported (Figs.

2 and 3). N. lenguarum occurs in the Chaco of Bolivia and

Paraguay, and possibly in the Argentinean Chaco as well

(Galliari et al. 1996; J. P. Jayat, in litt.). Our assignment of the

eastern Bolivian and Paraguayan specimens (TK 62259) to N.lenguarum must be regarded with caution for the following

reasons: the validity of N. lenguarum was not properly tested

because we failed to analyze topotype specimens from

Waikthlatingwaialwa; the validity of tapirapoanus, a taxon

currently synonymized under N. lenguarum, was not tested,

although the type locality of this taxon is less than 400 km from

eastern Bolivia (localities 37 and 41); specimens of Necromyscollected in the general area of Waikthlatingwaialwa (localities

34 and 35) are all referable to N. lasiurus; and our only

specimen referable to lenguarum from Paraguay comes from

Laguna Placenta and has yet to be compared with the holotype

of lenguarum. Until these issues are resolved, the taxonomic

status of N. lenguarum will remain elusive. However, our

analyses confirm the identity of an additional species of

Necromys in the Chaco region.

Necromys lasiurus as a senior synonym to N. benefactus andN. temchuki.—One of the most relevant results of our study is

that N. lasiurus, a broadly distributed species, appears para-

phyletic with respect to 2 southern species, N. benefactus and

N. temchuki. Haplotypes of N. benefactus do not form

June 2008 785D’ELIA ET AL.—SYSTEMATICS OF NECROMYS (SIGMODONTINAE)

a monophyletic group, being related to different haplotypes of

N. lasiurus. Haplotypes of N. temchuki sensu stricto, which

come from 3 geographically proximate localities (including the

type locality) in Misiones, Argentina, are very similar to each

other and form a well-supported clade (JK¼ 96, PP¼ 1.00). The

haplotype assignable to N. t. liciae is not sister to the N. t.temchuki clade.

The lack of monophyly for N. benefactus, N. lasiurus, and N.temchuki may simply reflect a disagreement between the inferred

gene tree and the species tree. Random fixation of alternative

haplotypes via lineage sorting may result in a pattern like the one

presented here (Neigel and Avise 1986). This also may be a case

of introgression via hybridization of the mitochondrial genome

of 1 species into the nuclear background of a 2nd (e.g., Patton and

Smith 1994). Currently, however, there is no reason to assume

that the recovered gene tree does not reflect the species tree. We

support this assertion on 2 lines of reasoning. First, there is no

alternative data set amenable to comparison with the mitochon-

drial tree presented here; that is, no other Necromys data set has

been analyzed with an explicit historical approach that suggests

that the mitochondrial tree departs from the species tree. There-

fore, the gene tree should be accepted as the best available hypoth-

esis of taxonomic relationships (Brower et al. 1996). Second, the

3 species are phenotypically very similar and no discrete

morphological groups can be delimited (Galliari and Pardinas

2000; Macedo and Mares 1987; G. D’Elıa et al., in litt.); similarly,

cytogenetic evidence shows that N. lasiurus and N. temchukihave the same chromosome number and morphology as well as

the same C- and G-banding patterns (Vitullo et al. 1986).

The phenotypic similarity of N. temchuki, a poorly described

species of northwestern Argentina, to N. lasiurus was already

noted by several authors, some of which questioned its specific

identity (e.g., Galliari et al. 1996). Taken at face value, our

results suggest that N. benefactus, N. lasiurus, and N. temchukirepresent a single biological species. Therefore, we formally

synonymyze N. benefactus and N. temchuki under N. lasiurus,

the oldest available name for this clade. Thus, the type species

of Necromys, the fossil form N. conifer, is hereby also relegated

to a junior synonym of N. lasiurus.

As here understood, N. lasiurus has one of the largest

distributions within the Sigmodontinae. Geographic limits are

not totally clear, but we have specimens assigned to this species

from the Ilha do Arapiranga in the Brazilian state of Para

(approximately 18199S), from southern Buenos Aires Province

(38.78S), from close to the Brazilian Atlantic coast, and from

the Sierra de Ambato on the Argentinean Andes. Although our

sampling of N. lasiurus is extensive (33 specimens and 25

populations), large areas (especially central Brazil) within these

limits are not represented in our analyses, and many others are

represented by singletons. Nevertheless, a preliminary picture

of geographic structure and demographic history of N. lasiuruscan now be articulated.

Observed divergence ranges from 0% to 3.9% and from

1.3% to 6.3% at the intra- and interpopulation level, respec-

tively. These are large values, and the last one is larger than

some values corresponding to comparisons between haplotypes

of N. lenguarum and N. lasiurus (range 5.2–8.1%); however,

these species’ reciprocal monophyly is strongly supported

(Figs. 2 and 3). Within lasiurus we note 2 clades, 1 composed

of 4 specimens from localities as far from each other as the Ilha

do Arapiranga (in the Brazilian state of Para) and 2 Paraguayan

localities, and another that includes the remaining individuals

sampled including 1 specimen from the type locality for

lasiurus (Lagoa Santa, Minas Gerais, Brazil). Observed

divergence between haplotypes from both clades ranges from

3.2% to 6.3%. Support for both clades is weak: JK ¼ 62, BS ¼1, PP ¼ 0.67 and JK ¼ 84, BS ¼ 4, PP ¼ 0.66, respectively.

Remarkably, despite its large geographic distribution, N.lasiurus does not show significant phylogeographic structure;

for example, the 2 main clades of N. lasiurus overlap in

Paraguay, whereas different haplotypes recovered in 1 collect-

ing locality at Serranıas de San Luis National Park (locality

32), fall into the 2 main clades of N. lasiurus. Several other

examples of this lack of phylogeographic structuring are

evident.

When all haplotypes of N. lasiurus are analyzed together,

both Tajima’s (1989) and Fu’s (1997) tests of neutrality

provide negative and significant values (D ¼ �1.59, P ¼ 0.03;

Fs ¼ �12.20, P ¼ 0.002), indicating that N. lasiurus may have

recently experienced an increase in population size. In

combination with the lack of phylogeographic structure, these

results appear to indicate that N. lasiurus has recently invaded

several regions of the current distributional range. In addition,

given the relatively large observed genetic divergence (up to

6.3%) between some pairs of haplotypes of N. lasiurus, it is

tempting to suggest that the presumptive range expansion

occurred either from a past population that was both large and

stable or from .1 population. The fossil record of N. lasiurusshows that in the course of a few hundred years the

distributional range of this species has shifted markedly

(Galliari and Pardinas 2000; Pardinas 1999). This type of

dynamic scenario may produce the lack of equilibrium between

genetic drift and gene flow that we observe in N. lasiurus.

Further tests of these hypotheses require the analyses of more

individuals and of a locus not linked to the cytochrome-b gene.

A better understanding of the history of tropical and subtropical

open areas of South America is needed to compose a detailed

scenario for the dynamic of N. lasiurus.

The geographic origin of Necromys.—Necromys, in spite of

its moderate specific diversity, is among the most widely

distributed genera within the sigmodontine radiation. Its

geographic distribution encompasses large expanses of open

environments both south and north of the Amazon basin. In

terms of its general distribution, Necromys resembles the rodent

genus Calomys (contra Almeida et al. 2007:460); the main

difference between them is that the latter reaches Patagonia in

southern South America, whereas Necromys has a larger

Andean distribution.

Despite their similar geographic distributions, it appears that

Calomys and Necromys have differerent biogeographic

histories. The main difference involves the reciprocal mono-

phyly of the groups of lowland and highland species of

Calomys (although a lowland clade was not recovered in the

maximum-parsimony analysis of Almeida et al. [2007]). This

786 JOURNAL OF MAMMALOGY Vol. 89, No. 3

implies a single split between lowland and Andean species,

probably mediated by the uplift of the Andes. Lowland species

of Calomys south and north of the Amazonia would have

differentiated after the Amazonian forest expanded (Almeida

et al. 2007; Salazar-Bravo et al. 2001). We can rule out

a similar scenario for Necromys because N. urichi from the

South American grasslands north of the Amazon is not sister to

species from the southern lowlands (Figs. 2 and 3). In fact, N.urichi is most closely related to N. amoenus, and, therefore, we

can rule out a single split between lowland and highland

species in Necromys. Incorporation of N. punctulatus in future

analyses may require a revision of this interpretation.

The oldest known Necromys corresponds to N. bonapartei�from late Pliocene sediments (approximately 4–3.5 million

years ago) from southern Buenos Aires Province, Argentina

(Pardinas and Tonni 1998). From slightly younger beds of the

same area Reig (1978) described Dankomys�, an extinct genus

morphologically resembling Necromys. To date, there is no

a formal phylogenetic analysis involving Dankomys�, but if

Dankomys� and Necromys are closely related, it would imply

a Pampean radiation of Necromys-like sigmodontines in the

late Pliocene, which in turn may suggest a Pampean origin for

Necromys (see Pardinas 1995). However, we note that a clade

of southern lowland species sister to the remaining species of

Necromys was never recovered in our analyses. In addition, we

note that none of our analyses recovered a southern clade that

was paraphyletic with respect to the highland species, as the

hypothesis of a southern origin for Necromys would predict.

Clearly, additional studies are needed to assess the historical

biogeography of Necromys. The discoveries of new fossils that

constitute conclusive evidence of the presence of a taxon in

a specific geologic time and place would be key evidence. In

addition, and as a 1st step toward the clarification of the place

of the origin of Necromys, a fully resolved and well-supported

topology at the base of the Necromys clade is much needed.

RESUMEN

Presentamos el estudio mas amplio a la fecha sobre la

sistematica de Necromys, un genero de roedores distribuido en

las areas abiertas al norte y sur de la Amazonıa y en los

pastizales andinos. El estudio esta basado en secuencias del gen

que codifica para el citocromo b; las mismas fueron analizadas

mediante analisis de maxima parsimonia y bayesiano. Los

analisis se efectuaron sobre secuencias de 62 especimenes de 51

localidades de Argentina, Bolivia, Brasil, Colombia, Paraguay,

Peru, Uruguay y Venezuela. Excepto una, las restantes especies

actualmente consideradas como integrantes del genero estan

representadas en el estudio. Necromys fue recobrado monofiletico

con una politomıa basal que involucra a 3 linajes. Uno de estos

linajes esta formado por la especie andina N. lactens, que ademas

presento una marcada estructura filogeografica. El segundo clado

esta formado por N. urichi de los pastizales del norte de America

del Sur y N. amoenus de los Andes centrales. Los resultados

sugieren que cada uno de estos taxones podrıa estar constituido

por mas de una especie biologica. El tercer clado esta formado

por las especies de las tierras bajas al sur de la Amazonıa. Dentro

de este clado, N. obscurus es hermana de las restantes especies.

Los haplotipos recobrados en especimenes asignados a N.benefactus, N. temchuki y N. lasiurus forman un clado, pero

estos taxones no forman grupos recıprocamente monofileticos ni

el clado muestra estructura geografica. Estos resultados genea-

logicos, discutidos en el contexto de la variacion genetica, son la

base de consideraciones taxonomicas (e.g., N. benefactus y N.temchuki son considerados sinonimos subjetivos de N. lasiurus) y

biogeograficas.

ACKNOWLEDGMENTS

We acknowledge the following institutions for access to tissues and

specimens: Museum of Vertebrate Zoology (Berkeley, J. Patton),

Museum of Texas Tech University (Lubbock, R. J. Baker, R. Owen,

and H. Garner), National Museum of Natural History (Washington,

D.C., L. Emmons), University of New Mexico’s Division of Genetic

Resources, Museum of Southwestern Biology (Albuquerque, T. Yates

and C. Parmenter), Division de Mastozoologıa, Coleccion Boliviana

de Fauna (La Paz, J. Vargas), Museum National d’Histoire Naturelle

(Paris, M. Tranier), and Natural History Museum (London, P.

Jenkins). Additional tissue samples were provided by L. Geise, E.

Hingst-Zaher, D. Birochio, R. Gonzalez Ittig, and J. Marinho. We

thank R. Monk (American Museum of Natural History, New York) for

providing photos of the holotype of A. urichi and R. Harbord and

P. Jenkis for providing photos of the holotypes of A. orbus and

B. negrito. Two anonymous reviewers made valuable comments on an

earlier version of this study. Financial support was provided by the

Direccion de Investigacion de la Universidad de Concepcion (DIUC

206.113.073-1.0) to GD, Consejo Nacional de Investigaciones

Cientıficas y Tecnicas (CONICET) to UFJP, National Science

Foundation (INT-9417252) and Faculty Research Award from Texas

Tech University to JSB, and Fundacion ProYungas to JPJ.

LITERATURE CITED

ALFARO, M. E., AND M. T. HOLDER. 2006. The posterior and the prior

in Bayesian phylogenetics. Annual Review of Ecology, Evolution,

and Systematics 37:19–42.

ALMEIDA, F. C., C. R. BONVICINO, AND P. CORDEIRO-ESTRELA. 2007.

Phylogeny and temporal diversification of Calomys (Rodentia,

Sigmodontinae): implications for the biogeography of an endemic

genus of the open/dry biomes of South America. Molecular

Phylogenetics and Evolution 42:449–466.

AMEGHINO, F. 1889. Contribucion al conocimiento de los mamıferos

fosiles de la Republica Argentina. Actas de la Academia Nacional

de Ciencias (Cordoba) 6:1–1027, atlas, 98 pls.

ANDERSON, S., AND N. OLDS. 1989. Notes on Bolivian mammals. 5.

Taxonomy and distribution of Bolomys (Muridae, Rodentia).

American Museum Novitates 2935:1–22.

ANDERSON, S., AND T. L. YATES. 2000. A new genus and species of

phyllotine rodent from Bolivia. Journal of Mammalogy 81:18–36.

BREMER, K. 1994. Branch support and tree stability. Cladistics

10:295–304.

BROWER, A. V. Z., R. DE SALLE, AND A. VOGLER. 1996. Gene trees,

species trees and systematics: a cladistic perspective. Annual

Review of Ecology and Systematics 27:423–450.

CABRERA, A. 1961. Catalogo de los mamıferos de America del Sur.

Revista del Museo Argentino de Ciencias Naturales ‘‘Bernardino

Rivadavia,’’ Zoologıa 4:xxii þ 309–732.

CABRERA, A. L. 1976. Regiones fitogeograficas argentinas. Pp. 1–85 in

Enciclopedia Argentina de agricultura y jardinerıa, tomo II,

June 2008 787D’ELIA ET AL.—SYSTEMATICS OF NECROMYS (SIGMODONTINAE)

fascıculo 1 (L. R. Parodi, ed.). Editorial Acme, Buenos Aires,

Argentina.

D’ELIA, G. 2003. Phylogenetics of Sigmodontinae (Rodentia,

Muroidea, Cricetidae), with special reference to the akodont group,

and with additional comments on historical biogeography. Cladis-

tics 19:307–323.

D’ELIA, G., E. M. GONZALEZ, AND U. F. J. PARDINAS. 2003.

Phylogenetic analysis of sigmodontine rodents (Muroidea), with

special reference to the akodont genus Deltamys. Mammalian

Biology 68:351–364.

D’ELIA, G., AND U. F. J. PARDINAS. 2004. Systematics of Argentinean,

Paraguayan, and Uruguayan swamp rats of the genus Scapteromys(Rodentia, Cricetidae, Sigmodontinae). Journal of Mammalogy

85:897–910.

D’ELIA, G., AND U. F. J. PARDINAS. 2007. Putting names to the

phylogenetic diversity of neotropical sigmodontine rodents: new

genera for known species. Mammalia 71:146–149.

D’ELIA, G., U. F. J. PARDINAS, P. TETA, AND J. L. PATTON. 2007.

Definition and diagnosis of a new tribe of sigmodontine rodents

(Cricetidae: Sigmodontinae), and a revised classification of the

subfamily. Gayana 71:187–194.

DA SILVA, M. N., AND J. L. PATTON. 1993. Amazonian phylogeog-

raphy: mtDNA sequence variation in arboreal echimyid rodents

(Caviomorpha). Molecular Phylogenetics and Evolution 2:243–255.

DNASTAR. 2003. Lasergene: expert sequence analysis software.

Version 5. DNASTAR, Madison, Wisconsin.

EXCOFFIER, L., P. E. SMOUSE, AND J. M. QUATTRO. 1992. Analysis of

molecular variance inferred from metric distances among DNA

haplotypes—application to human mitochondrial-DNA restriction

data. Genetics 131:479–491.

FARRIS, J. S. 1982. The logical basis of phylogenetic analysis. Pp. 7–

36 in Advances in cladistics: proceedings of the second meeting of

the Willi Hennig Society (N. Planick and V. Funk, eds.). Columbia

University Press, New York.

FARRIS, J. S., V. A. ALBERT, M. KALLERSJO, D. LIPSCOMB, AND A. G.

KLUGE. 1996. Parsimony jackknifing outperforms neighbor-joining.

Cladistics 12:99–124.

FU, Y. X. 1997. Statistical tests of neutrality of mutations against

population growth, hitchhiking and background selection. Genetics

147:915–925.

GALEWSKI, T., M. TILAK, S. SANCHEZ, P. CHEVRET, E. PARADIS, AND

E. J. P. DOUZERY. 2006. The evolutionary radiation of Arvicolinae

rodents (voles and lemmings): relative contribution of nuclear and

mitochondrial DNA phylogenies. BMC Evolutionary Biology 6:

article 80.

GALLIARI, C. A., AND U. F. J. PARDINAS. 2000. Taxonomy and

distribution of the sigmodontine rodents of genus Necromys in

central Argentina and Uruguay. Acta Theriologica 45:211–232.

GALLIARI, C. A., U. F. J. PARDINAS, AND F. GOIN. 1996. Lista comentada

de los mamıferos argentinos. Mastozoologıa Neotropical 3:39–61.

GANNON, W. L., R. S. SIKES, AND THE ANIMAL CARE AND USE

COMMITTEE OF THE AMERICAN SOCIETY OF MAMMALOGISTS. 2007.

Guidelines of the American Society of Mammalogists for the use of

wild mammals in research. Journal of Mammalogy 88:809–823.

GYLDENSTOLPE, N. 1932. A manual of neotropical sigmodont rodents.

Kunglia Svenska Vetenskapsakademiens Handlingar, Stockholm.

Series 3, 11:1–164.

HALLOY, S. 1978. Contribucion al estudio de la zona de Huaca-Huasi,

Cumbres Calchaquıes, (Tucuman, Argentina) I: la biota actual en

relacion con la geologıa historica. Seminario de Licenciatura,

Facultad de Ciencias Naturales e Instituto Miguel Lillo, Universidad

Nacional de Tucuman, Tucuman, Argentina.

HUELSENBECK, J. P., F. RONQUIST, R. NIELSEN, AND J. P. BOLLBACK.

2001. Bayesian inference of phylogeny and its impact on

evolutionary biology. Science 294:2310–2314.

JAYAT, J. P., AND S. PACHECO. 2006. Distribucion de Necromys lactens

y Phyllotis osilae (Rodentia: Cricetidae: Sigmodontinae) en el

noroeste argentino: modelos predictivos basados en el concepto de

nicho ecologico. Mastozoologıa Neotropical 13:69–88.

KLUGE, A. J., AND J. S. FARRIS. 1969. Quantitative phyletics and the

evolution of anurans. Systematic Zoology 18:1–32.

LINARES, O. J. 1998. Mamıferos de Venezuela. Sociedad Conserva-

cionista Audobon de Venezuela, Caracas, Venezuela.

MACEDO, R. H., AND M. A. MARES. 1987. Geographic variation in the

South American cricetine rodent Bolomys lasiurus. Journal of

Mammalogy 68:578–594.

MASSOIA, E., AND A. FORNES. 1967. El estado sistematico, distribucion

geografica y datos etoecologicos de algunos mamıferos neotropicales

(Marsupialia y Rodentia) con la descripcion de Cabreramys, genero

nuevo (Cricetidae). Acta Zoologica Lilloana 23:407–430.

MASSOIA, E., AND U. F. J. PARDINAS. 1993. El estado sistematico de

algunos muroideos estudiados por Ameghino en 1889. Revalida-

cion del genero Necromys (Mammalia, Rodentia, Cricetidae).

Ameghiniana 30:407–418.

MUSSER, G. G., AND M. D. CARLETON. 2005. Superfamily Muroidea.

Pp. 894–1531 in Mammal species of the world: a taxonomic and

geographic reference (D. E. Wilson and D. M. Reeder, eds.). 3rd ed.

Johns Hopkins University Press, Baltimore, Maryland.

NEIGEL, J. E., AND J. C. AVISE, 1986. Phylogenetic relationships of

mitochondrial DNA under various demographic models of

speciation. Pp. 515–534 in Evolutionary processes and theory (S.

Karlin and E. Nevo, eds.). Academic Press, San Diego, California.

NELSON, G. 1979. Cladistic analysis and synthesis: principles and

definitions, with a historical note on Adanson’s Familles des

Plantes (1763–1764). Systematic Zoology 28:1–21.

OLIVEIRA, J. A. 1992. Estrutura da variacao craniana em populac~oes de

Bolomys lasiurus (Lund, 1841) (Rodentia: Cricetidae) do nordeste

do Brasil. Coordenacao de Pos-Graduacao em Zoologia, Universi-

dade Federal do Rio de Janeiro, Rio de Janeiro, Brazil.

ORTIZ, P. E., U. F. J. PARDINAS, AND J. E. POWELL. 1998. Roedores del

Pleistoceno–Holoceno del Valle de Tafı (Tucuman, Argentina).

Aspectos sistematicos y paleoambientales. Acta Geologica Lilloana

18:173–174.

ORTIZ, P., U. F. J. PARDINAS, AND S. STEPPAN. 2000. A new fossil

phyllotine (Rodentia: Muridae) from northwestern Argentina and the

relationships of Reithrodon group. Journal of Mammalogy 81:37–51.

PARDINAS, U. F. J. 1995. Sobre las vicisitudes de los generos

Bothriomys Ameghino, 1889, Euneomys Coues, 1874 y Graomys

Thomas, 1916 (Mammalia, Rodentia, Cricetidae). Ameghiniana

32:173–180.

PARDINAS, U. F. J. 1999. Fossil murids: taxonomy, paleoecology, and

paleoenvironments. Quaternary of South America and Antarctic

Peninsula 12:225–254.

PARDINAS, U. F. J., AND E. TONNI. 1998. Procedencia estratigrafica y

edad de los mas antiguos muroideos (Mammalia, Rodentia) de

America del Sur. Ameghiniana 35:473–475.

PATTON, J. L., AND M. F. SMITH. 1994. Paraphyly, polyphyly and the

nature of species boundaries in pocket gophers (genus Thomomys).

Systematic Biology 43:11–26.

PROVENSAL, M. C., G. E. CALDERON, M. B. CHIAPPERO, C. N.

GARDENAL, J. J. POLOP, AND M. S. SABATTINI. 2005. Morphometric

and allozymic characterization of Necromys benefactus populations

in central Argentina. Mastozoologıa Neotropical 12:261–268.

788 JOURNAL OF MAMMALOGY Vol. 89, No. 3

REIG, O. A. 1978. Roedores cricetidos del Plioceno superior de la

provincia de Buenos Aires (Argentina). Publicaciones del Museo

Municipal de Ciencias Naturales de Mar del Plata, ‘‘Lorenzo

Scaglia’’ 2:164–190.

REIG, O. A. 1986. Diversity patterns and differentiation of high

Andean rodents. Pp. 404–439 in High altitude tropical biogeogra-

phy (F. Vuilleumier and M. Monasterio, eds.). Oxford University

Press, New York.

REIG, O. A. 1987. An assessment of the systematics and evolution of

the Akodontini, with the description of new fossil species of

Akodon (Cricetidae, Sigmodontinae). Pp. 347–399 in Studies in

neotropical mammalogy: essays in honor of Philip Hershkovitz

(B. D. Patterson and R. M. Timm, eds.). Fieldiana: Zoology (New

Series) 39:viii þ 1–506.

ROHMEDER, G. 1942. La glaciacion diluvial de los Nevados del

Aconquija (parte austral). Monografıas del Instituto de Estudios

Geograficos 2. Universidad Nacional de Tucuman, Tucuman,

Argentina.

RONQUIST, F., AND J. P. HUELSENBECK. 2003. MRBAYES 3: Bayesian

phylogenetic inference under mixed models. Bioinformatics

19:1572–1574.

SALAZAR-BRAVO, J., J. W. DRAGOO, D. S. TINNIN, AND T. L. YATES.

2001. Phylogeny and evolution of the neotropical rodent genus

Calomys: inferences from mitochondrial DNA sequence data.

Molecular Phylogenetics and Evolution 20:173–184.

SCHNEIDER, S., D. ROESSLI, AND L. EXCOFFIER. 2000. Arlequin, version

2.000: a software for population genetics data analysis. Genetics

and Biometry Laboratory, University of Geneva, Geneva, Switzerland.

SIMMONS, M. P., K. M. PICKETT, AND M. MIYA. 2004. How meaningful

are Bayesian support values? Molecular Biology and Evolution

21:188–199.

SMITH, M. F., AND J. L. PATTON. 1991. Variation in mitochondrial

cytochrome b sequence in natural populations of South American

akodontine rodents (Muridae: Sigmodontinae). Molecular Biology

and Evolution 8:85–103.

SMITH, M. F., AND J. L. PATTON. 1993. The diversification of South

American murid rodents: evidence from mitochondrial DNA

sequence data for the akodontine tribe. Biological Journal of the

Linnean Society 50:149–177.

SMITH, M. F., AND J. L. PATTON. 1999. Phylogenetic relationships and

the radiation of sigmodontine rodents in South America: evidence

from cytochrome b. Journal of Mammalian Evolution 6:89–128.

SORENSON, M. D. 1999. TreeRot, version 2. Boston University,

Boston, Massachusetts.

STEPPAN, S. J., R. M. ADKINS, P. Q. SPINKS, AND C. HALE. 2005.

Multigene phylogeny of the Old World mice, Murinae, reveals

distinct geographic lineages and the declining utility of mitochon-

drial genes compared to nuclear genes. Molecular Phylogenetics

and Evolution 37:370–388.

SWOFFORD, D. 2000. PAUP*: phylogenetic analysis using parsimony

(* and other methods), 4.0. Sinauer Associates, Inc., Publishers,

Sunderland, Massachusetts.

TAJIMA, F. 1989. The effect of change in population size on DNA

polymorphism. Genetics 123:597–601.

TATE, G. H. H. 1932. The taxonomic history of the South and Central

American akodont rodent genera: Thalpomys, Deltamys, Thapto-

mys, Hypsimys, Bolomys, Chroeomys, Abrothrix, Scotinomys,

Akodon (Chalcomys and Akodon), Microxus, Podoxymys, Lenoxus,

Oxymycterus, Notiomys, and Blarinomys. American Museum

Novitates 582:1–32.

THOMAS, O. 1916. The grouping of the South-American Muridae

commonly referred to Akodon. Annals and Magazine of Natural

History, Series 8, 18:336–340.

THOMPSON, J. D., T. J. GIBSON, F. PLEWNIAK, F. JEANMOUGIN, AND D. G.

HIGGINS. 1997. The Clustal X Windows interface: flexible strategies

for multiple sequence alignment aided by quality analysis tools.

Nucleic Acids Research 24:4876–4882.

VENTURA, J., M. J. LOPEZ-FUSTER, M. SALAZAR, AND R. PEREZ-

HERNANDEZ. 2000. Morphometric analysis of some Venezuelan

akodontine rodents. Netherlands Journal of Zoology 50:487–501.

VITULLO, A. D., ET AL. 1986. Cytogenetics of South American akodont

rodents (Cricetidae): new karyotypes and chromosomal banding

patterns of Argentinean and Uruguayan forms. Journal of

Mammalogy 67:69–80.

VOSS, R. S. 1991. An introduction to neotropical genus Zygodontomys.

Bulletin of the American Museum of Natural History 210:1–113.

Submitted 14 August 2007. Accepted 28 November 2007.

Associate Editor was Carey W. Krajewski.

APPENDIX IThe specimens of Necromys, Akodon, Deltamys, Thalpomys, and

Thaptomys used in the present study are listed here. Accession

numbers are indicated for those specimens whose sequences were

retrieved from GenBank. See Fig. 1 for locality numbers and locations

of sites. Museum and collection acronyms and personal field numbers

are as follows. Argentina: Coleccion de Mamıferos del Centro

Nacional Patagonico, Puerto Madryn (CNP); Museo de Ciencias

Naturales y Tradicional de Mar del Plata, Mar del Plata (MMP); field

number of J. Pablo Jayat (JPJ, vouchers will be deposited in Coleccion

Mamıferos Lillo, Tucuman). Bolivia: VCC field number of Veronica

Chavez (VCC, vouchers will be deposited in Museo de Historia

Natural Noel Kempff Mercado, Santa Cruz). Brazil: Museu Nacional,

Rio do Janeiro (MN); field number of Lena Geise and field crew (CD,

voucher will be deposited at the Museu de Zoologia Universidade de

Sao Paulo, Brazil), MZUSP; field and laboratory numbers obtained

from Erika Hingst-Zaher (BlasLS052III, IA to be deposited at the

Museu Nacional, Brazil); field number of Jorge Marinho (JR, voucher

will be deposited at Colecao Zoologica da Universidade Regional de

Blumenau, Brazil). France: Museum National d’Histoire Naturelle,

Paris (MNHN). United Kingdom: The Natural History Museum,

London (BMNH). United States: American Museum of Natural

History, New York (AMNH); Museum of Southwestern Biology,

Albuquerque (MSB and NK: tissue accession prefix); Museum of

Vertebrate Zoology, University of California, Berkeley (MVZ); The

Museum, Texas Tech University, Lubbock (TK: tissue accession

prefix); University of Michigan Museum of Zoology, Ann Arbor

(UMMZ); National Museum of Natural History, Washington, D.C.

(USNM). Uruguay: field number of Guillermo D’Elıa (GD, vouchers

will be deposited at Museo Nacional de Historia Natural y

Antropologıa, Montevideo); field numbers of the Seccion Evolucion,

Facultad de Ciencias, Universidad de la Republica (EV, vouchers will

be deposited at Nacional de Historia Natural y Antropologıa,

Montevideo). An asterisk (*) or a pound symbol (#) indicates that

before this study the specimen in question was assigned to N.benefactus or N. temchuki, respectively; both taxa are here regarded as

junior synonyms of N. lasiurus. gb indicates that the sequence was

retrieved from GenBank; all others were generated in this study.

Necromys amoenus.—ARGENTINA: 1) Salta, Santa Victoria, 1

km ENE de Rodeo Pampa, km 59 de Ruta Provincial No. 7, 3,080 m,

22814947.70S, 658394.30W (JPJ 1318: EF531643). BOLIVIA: 2)

June 2008 789D’ELIA ET AL.—SYSTEMATICS OF NECROMYS (SIGMODONTINAE)

Tarija, Serrania del Sama, 3,200 m, 218279S, 648529W (MSB 67194:

AF159283 gb). PERU: 3) Cuzco, 20 km N (by road) Paucartambo, km

100, 3,580 m, 1381299.70S, 7084090.00W (MVZ 171563: AY273911

gb). 4) Puno, 12 km S Santa Rosa, 3,960 m, 1484399.40S, 7084798.90W

(MVZ 172878: M35711 gb; MVZ 172879: M35712 gb).

Necromys lactens.—ARGENTINA: 5) Catamarca, Ambato, Las

Chacritas, aprox. 28 km al NNW de Singuil, sobre Ruta Provincial No.

1, 1,888 m, 27842924.20S, 65854940.60W (JPJ 552: EF531646); 6)

Catamarca, Valle Viejo, aprox. 2 km al SE de Huaico Hondo, sobre

Ruta Provincial No. 42, al E del Portezuelo, 1,992 m, 28825910.90S,

65832940.50W (JPJ 419: EF531644); 7) Jujuy, Santa Barbara, La

Antena, Sierra del Centinela, al S de El Fuerte, 2,350 m,

24817956.460S, 6482399.30W (JPJ 953: EF531649); 8) Jujuy, Tum-

baya, Barcena, aprox. 3 km al S, sobre Ruta Nacional No. 9, 1,808 m,

2480920S, 65826951.60W (JPJ 630: EF531647); 9) Salta, Oran, Abra de

Cienaga Negra, aprox. 3 km al SE, 3,090 m, 238199490S, 648539320W

(JPJ 721: EF531648); 10) Salta, Santa Victoria, 1 km ENE de Rodeo

Pampa, km 59 de Ruta Provincial No. 7, 3,080 m, 22814947.70S,

658394.30W (JPJ 1315: EF531650); 11) Tucuman, Trancas, aprox. 10

km al S de Hualinchay, sobre el camino a Lara, 2,300 m,

26819920.20S, 65836945.50W (JPJ 447: EF531645).

Necromys lasiurus.—ARGENTINA: 12) Buenos Aires, Bahıa

Blanca, Bahıa Blanca, 388439S, 628169W (CNP 519*: EF531654;

CNP 520*: EF531655); 13) Buenos Aires, Campamento Base del

Cerro Ventana, 388049S, 628019W (BM 79.1665*: EF531651; CNP

472*: EF531652); 14) Buenos Aires, San Mateo, 388399340S,

60859930.60W (CNP 473*: EF531653); 5) Catamarca, Ambato, Las

Chacritas, aprox. 28 km al NNW de Singuil, sobre Ruta Provincial

No. 1, 1,888 m, 27842924.20S, 65854940.60W (JPJ 530: EF531661);

15) Formosa, 17 km W-SW de Colonia Mayor Villafane, 268129570S,

598159380W (CNP 533#: AY273913 gb); 16) Misiones, Estancia

Santa Ines, 278319320S, 558529190W (CNP 534#: EF531658; CNP

535#: AY273914 gb); 17) Misiones, Leandro N. Alem, 278369S,

558199W (CNP 795#: EF531659); 18) Misiones, Santa Ana, 278239S,

558359W (CNP 801#: EF531660); 19) Santa Fe, Berna, 298169S,

598529W (CNP 531*: EF531656; CNP 532*: EF531657). BRASIL:

20) Bahia, Lencois, Chapada Diamantina, Remanso, 509 m,

12836941.50S, 41821950.90W (CD 21: EF531689); 21) Bahia, Mucuge,

Parque Nacional da Chapada Diamantina, Rio Cumbuca, 1,002 m,

13829410S, 418209580W (MN 69862: EF531693); 22) Minas Gerais,

Caratinga, Fazenda Montes Claros, 198509S, 418509W (MN 48052:

EF531690); 23) Minas Gerais, Lagoa Santa, 198399S, 438549W

(LS052III: EF531688); 24) Para, Ilha de Arapiranga, 0 m, 18209S,

488349W (IA01: EF531691; IA042: EF531692); 25) Rio Grande do

Sul, Rondinha, Parque Estadual Florestal de Rondinha, 28844919.740S,

50816940.570W (JR 346: EF531662); 26) Sao Paulo, Tupi Paulista,

218229S, 518369W (NK 42164: EF531663). PARAGUAY: 27) Alto

Paraguay, Palmar de Las Islas, 19837947.30S, 608369450W (TK 65362:

EF531673); 28) Amambay: Bella Vista, Colonia Sargento Dure, 3 km E

by road of Rio Apa, 228109S, 56825915.240W (NK 27506: EF531664;

NK 27519: EF531665); 29) Canindeyu, Reserva Natural del Bosque

Mbaracayu, 24808.939S, 558189W (TK 61813: EF531666); 30)

Canindeyu, Reserva Natural del Bosque Mbaracayu, 24809.619S,

55816.999W (TK 63980: EF531668; TK 63982: EF531669); 31)

Canindeyu, Reserva Natural Privada Itabo, 24827.689S, 54838.339W

(TK 63511: EF531667); 32) Concepcion, Parque Nacional Serranıa de

San Luis, 22835.959S, 57821.289W (TK 64242: EF531670; TK 64268:

EF531671); 33) Concepcion, Parque Nacional Serranıa de San Luis,

22840.969S, 57821.529W (TK 64302: AY273912 gb); 34) Presidente

Hayes, 8 km NE Juan de Zalazar, 2386900S, 59818900W (UMMZ

134431: U03528 gb); 35) Presidente Hayes, Estancia Loma Pora,

2383295.30S, 57833938.30W (TK 64472: EF531672).

Necromys lenguarum.—BOLIVIA: 36) Chuquisaca, Rıo Limon,

1,300 m, 198329510S, 64847958.50W (NK 21680: EF531677); 37)

Santa Cruz, Velasco, Parque Nacional Noel Kempff Mercado,

Campamento Los Fierros, 215 m, 148339400S, 608559400W (VCC

113: EF531680; USNM 584530: EF531679); 38) Santa Cruz, Estancia

San Marcos, 6 km W Ascencion, approx. 400 m, 15853911.40S,

63811910.390W (NK 102253: EF531678); 39) Santa Cruz, 6 km N

Buen Retiro, 300 m, 17812950.90S, 63837958.40W (NK 12151:

EF531674); 40) Santa Cruz, 4.5 km N, 1.5 km E Cerro Amboro,

Rio Pitisama, 620 m, 178449S, 638409W (NK 14011: EF531676); 41)

Santa Cruz, 4 km N, 1 km W of Santiago de Chiquitos, 700 m,

188179510S, 59835958.130W (NK12315: EF531675). PARAGUAY:

42) Alto Paraguay, Laguna Placenta, 70 m, 2188937.10S, 59824951.50W

(TK 62259: EF531681).

Necromys obscurus.—ARGENTINA: 43) Buenos Aires, 1 km

aguas arriba puente sobre RP 11, arroyo de las Brusquitas,

388149080S, 578469520W (CNP 891: EF531684; CNP 892:

EF531685); 44) Buenos Aires, Mar del Plata, 0 m, 388009S,

578339W (MMP 4008: EF531686). URUGUAY: 45) Montevideo,

Parque Lecoq, 34848.1219S, W 56820.4609W (GD 754: EF531683);

46) San Jose, Estancia El Relincho, 348209S, 568599W (EV 1028:

EF531682).

Necromys urichi.—COLOMBIA: 47) Paramo de Choachi, alrede-

dores de Bogota, 48339N, 738409W (MNHN 2880: EF531687).

VENEZUELA: 48) Amazonas, Cerro Neblina, 0851924.170N,

65857913.570W (USNM 560661: U03549 gb); 49) Bolivar, 5.2 km

NE San Ignacio Yuruani (AMNH 257281: AY273918 gb); 50) Sucre,

Finca Vuelta Larga, 9.7 km (by road) SE Guaraunos, 10827959.780N,

63859510W (AMNH 257287: AY273919 gb).

Akodon boliviensis.—PERU: Puno, 12 km S Santa Rosa [De

Ayaviri]; elevation 3,950 m (MVZ 171607: M35691 gb).

Deltamys kempi.—ARGENTINA: Buenos Aires, La Balandra,

348549070S, 578459560W (CNP 893: AY 195861 gb).

Thalpomys cerradensis.—BRAZIL: Piauı, Estacao Ecologica de

Urucui-Una (MZUSP 30400: AY273915 gb).

Thaptomys nigrita.—BRAZIL: Sao Paulo, Salesopolis; Estacao

Biologica de Boraceia, 3 km E, 28 km SE Biritiba-Mirim, 850 m

(MVZ 1830 44: AF108666 gb).

790 JOURNAL OF MAMMALOGY Vol. 89, No. 3